Ever since Seattle enacted a measure to raise its minimum wage to a nation-leading $15 an hour, conservatives have been sharpening their pencils to show the raise is a job-killer.
Mike Patton of Forbes broke out of the box early, with a piece on Sept. 28 asserting that "thus far the data doesn’t bode well for supporters of this law." His evidence was an apparent spike in Seattle's
"The unemployment rate will likely trend higher for several years as businesses seek ways to mitigate the negative financial consequences of this law," Patton wrote.
The issue isn't trivial. Dozens of municipalities are contemplating increases in their minimum wage or begun implementing long-term changes. Los Angeles will raise its minimum July 1 from the state-mandated $10 to $10.50, with the goal of $15 in 2020.
But the problem with using Seattle as an early warning signal for minimum wage increases is that, as yet, there's almost no good information. Patton, for instance, used a few months of non-seasonally adjusted data, but he may not have noticed that Seattle employment falls every April before recovering in subsequent months. Adjusting for seasonal variation, as the Bureau of Labor Statistics does routinely, the spike in the unemployment rate disappeared.
In fact, firm figures for city employment and wages in the post-April 1 period won't become available until early this year. A study team from the University of Washington, which has contracted with the city to analyze the effect, won't produce its first, very preliminary report, until mid-year.
That brings us to Mark J. Perry of the pro-business American Enterprise Institute, one of the most assiduous critics of minimum wage increases in general and Seattle's in particular. By my count, he's written about a dozen posts on his AEI blog wholly or partially about Seattle since the end of March, among more than 25 posts attacking minimum wage increases generally.
Perry has been looking hard for evidence of a decline in Seattle jobs, especially in the restaurant industry, for months. In August and again in October he claimed he found it: In the first piece, he reported that restaurants in the Seattle metropolitan statistical area had shed 1,300 jobs between January and June 2015, while employment in other categories had risen. In his follow-up extending the statistics through September, he reported that restaurant employment in the Seattle metropolitan statistical area had fallen by a net 700, while restaurant employment in the state, outside of Seattle, had risen.
The problem, as several commentators pointed out, was that the Seattle metropolitan statistical area is not the same as the city of Seattle. It's huge — three counties encompassing 39 cities, with a population of 3.6 million compared to Seattle's 660,000 residents — and only Seattle had enacted a $15 minimum wage. As regional economist Annaliese Vance-Sherman told the economics blogger Invictus of Ritholtz.com, "It is not possible to draw conclusions about the city based on the MSA."
Now, Perry is back, armed with what he says are Seattle-only statistics. "Seattle's 'radical experiment' might be a model for the rest of the nation not to follow," he wrote on Feb. 18. He cited figures from the Bureau of Labor Statistics showing that Seattle employment fell by more than 11,000 from April, the date of the first minimum wage hike, through December. He compared these numbers to the Seattle MSA, writing that "while jobs in the city of Seattle were tanking starting last April, employment in the suburbs surrounding Seattle was increasing steadily to a new record high in November."
Unfortunately, local economists say Perry is still using bad data. Although he attributes the city-only numbers to the Bureau of Labor Statistics, they're not reliable jobs numbers. Perry's source is the Local Area Unemployment Statistics file, or LAUS, which is based on a small sampling. It's aimed at counting the number of employed people living in the sample area (in this case, Seattle), not the number of jobs. The data are "prone to error," University of Washington economist Jacob Vigdor told me by email, and "basically worthless for any serious analysis."
Indeed, Vigdor — who is overseeing the university's analysis of minimum-wage data — notes that the same statistics for Bellevue and Everett, Wash., showed exactly the same percentage decrease that Perry found in Seattle, even though they haven't increased their minimum wage. (See below.)
"My best guess is that the BLS has produced the figures for all three cities using a tiny amount of data mixed with a large number of assumptions," he said.
Vigdor observes that it's simply too soon to come to any conclusions about the effect of the wage increase in the restaurant sector. Among other things, restaurants say that $11 an hour, the statutory minimum as of April 1, was the going rate for restaurant employees even before the law took effect. That means that employers actually didn't see an effective increase until Jan. 1, when the rate moved to $12 an hour for small employers and $13 for those with more than 500 workers.
Perry defends his analysis. "Perhaps those BLS data are unreliable," he told me by email, but he maintains that all the data still suggest there are "negative employment trends in the Seattle area." He adds, "We can not necessarily blame the minimum wage increases ... but that is one factor that has to be considered.
"The jury is still out on the $15 minimum wage," he said, "and it will take years to assess its impact. I'm simply pointing to some possible evidence in employment trends that might suggest that there is early evidence of some effects."
Yet it's not clear that there really is early evidence of "some effects," as opposed to flawed statistics that can be massaged to show it. Critics of the Seattle minimum wage are, at best, jumping the gun.
"I would not go so far as to say that the [Perry] study contradicts our conclusions," Vigdor said. "We don't have the data yet, there is nothing to contradict. ... Our primary interest is in ensuring that dialogue over Seattle's experience is conducted using reliable evidence, which the LAUS numbers are not."